Study On Automatic Parking System Based On Fuzzy Control

With the development of automobile industry and the improvement of people's living standards, the number of cars has increased year by year; accidents happen more often when cars are backing, which has got widespread attention. To solve the problem, people mostly install parking assistants such as radar and image display on the car to help parking cars. Although the assistant is useful for the driver to get environment information when parking cars, but not completely solve the problem because the driver still needs to park the car himself. In order to solve this problem fundamentally, the only way is to realize automatic car backing system which makes the entire car backing process complete automatically and intelligently.The automatic parking system is made up of three parts, including environmental data acquisition subsystem, the central controller and the steering control subsystem. Environmental data acquisition subsystem is responsible to collect vehicle position data, detect parking space and ranging distance between vehicle and obstacles, then transmit these data to central controller, to determine parking plan. After analyzing these collected data, the central controller will get the car's information of current position and the target location, and other surrounding parameters, then make automatic backup strategy based on the above parameters, and convert them into electrical signals. The steering control subsystem will carry out the instructions from central controller to make the car drive some distance or turn some angles after receiving the signals, to realize automatic parking in the end.This paper mainly do research on two parts:design a fuzzy controller based on car backing experience from experts and design a ultrasonic positioning system, to detect parking spaces and collect car's location data.Firstly, the paper analyzes the whole process of car backing, and then establishes the kinematics model instead of dynamic model because of low speed driving, the whole backing process follows Ackerman steering law. This paper then calculates and gains the feasible initial region where starting to back a car according to car's kinematics backing model, analyzes the situations including when the initial heading angle is zero and when the angle is not zero, which prepares for the controller designed subsequently. Then after studying theory of fuzzy logic control, the paper designs a controller contains three inputs and one output, sets the range for these inputs and output and their membership functions. Learning from experts and drivers who have experience of backing cars, this paper designs the fuzzy logic controller with Matlab based on above work, and simulate by Simulink, to testify efficiency of the controller.This paper designs an ultrasonic detection system, using ultrasonic sensors to detect parking spaces and locate the position of the car. This ultrasonic system uses DSP2812 as controller, the hardware design including DSP minimum system, ultrasonic transmit circuit, ultrasonic echo receive circuit and so on; the software part contains DSP initialization subroutine, interrupt service subroutine, ranging subroutine and so on. In the end, this paper completes ranging experiments and parking space detection experiments using the ultrasonic system, analyzes causes of the error and then proposes improvement measures...